Cancer is one of the leading causes of death in the US, and hepatocellular carcinoma (HCC) is one of the most lethal cancers. In 2015, HCC had an overall mortality rate of 69% and accounted for over 21,000 deaths in the US alone. Numerous mutations, including chromosome rearrangements, have been discovered in HCC. Many chromosome rearrangements are recurrent, found in many HCCs and other cancers. For instance, of 14 fusion genes (products of chromosome rearrangement) that we found in prostate cancers, some were found at significant frequencies in HCCs: 15.7% (11/70) tumors positive for MAN2A1-FER, 78.6% (55/70) for SLC45A2-AMACR, 12.9% (9/70) for TRMT11-GRIK2, 2.9% (4/70) for CLTC-ETV1, 2.9% (2/70) for DOCK7-OLR1, 84.3% (59/70) for ZMPSTE24-ZMYM4 and 82.9% (58/70) for Pten-NOLC1. MAN2A1-FER has been found to have constitutive tyrosine protein kinase activity, and was characterized to play a critical role in HCC development. In addition, SLC45A2-AMACR and Pten-NOLC1 were also found to be oncogenic to drive the liver cancer development. We recently developed an approach to treat human cancers using CRISPR-cas9 editing to insert a suicide gene at the chromosomal breakpoint of a fusion gene. Using this approach to target MAN2A1-FER and TMEM135-CCDC67, we achieved partial remission of tumors in mice with grafts of human liver and prostate cancers. Specifically, we designed one adenovirus to deliver the nickase Cas9D10A and gRNAs targeting the breakpoint sequences and another to deliver an EGFP-HSV1-thymidine kinase (EGFP-HSV1-tk) construct flanked by sequences homologous to sequences on either side of the breakpoint. Infection with both viruses resulted in breakpoint- dependent expression of EGFP-tk and ganciclovir-mediated apoptosis in cancer cells containing the breakpoint, but not in cells lacking the breakpoint. All mice with xenografts showed significant reductions of tumor burden with no mortality after 8 weeks of observation. In contrast, all control mice, including animals xenografted with human HCC line HUH7 but treated with incorrect gRNA or animals xenografted with HCC line HEP3B that lacks a MAN2A1-FER breakpoint and treated with MAN2A1-FER targeting reagents, died within 7 weeks of receiving a xenograft. Our results suggest that Cas9-mediated suicide gene insertion might be a highly specific and robust cancer gene therapy. Because chromosome rearrangements and mutations are present in the genomes of many human cancers, targeting these alterations for EGFP- HSV1-tk insertion may be an effective cancer treatment. Based on these findings, we hypothesize that targeting chromosomal breakpoints of fusion genes, somatic mutations in cancer cells or combination of both alterations is an effective approach to treat human cancers, including liver cancer. The specific aims are: 1) To determine whether genome targeting therapy is effective in treating liver cancers induced by fusion genes MAN2A1-FER, SLC45A2-AMACR and Pten-NOLC1; 2) To determine whether genome therapy is effective in targeting CTNNB1 mutation induced liver cancer in mice; And 3) To determine whether genome targeting therapy is adaptive and effective in treating DEN induced liver cancers without preconception of resident mutation.